Re: [Biomodels-net-support] BioPax format

[Nicolas Rodriguez]

On 14/06/18 13:31, Koehorst, Jasper wrote:
> Hello,
>
> We have developed genome annotation in RDF and were looking into expanding this to biopax for metabolic modelling.
> I was very happy to see that the biomodels are available as biopax and SBML. However I was wondering if the conversion tool (assuming that you have one) is available for the public?

Yes, the converters are part of SBFC (https://sf.net/p/sbfc) and are
available as a distribution and also online (at
http://www.ebi.ac.uk/biomodels/tools/converters).

Unfortunately, the latest version of the sbml to biopax which correct
several problems is not yet released but I have very recently produce a
zip file that contain it, you can download it there:
http://lenoverelab.org/perso/rodrigue/sbfc-1.3.8-SNAPSHOT.zip (We hope
to have a new official release of SBFC with several improvements and
fixes this summer).

Once you unzip the distribution somewhere, you can use the script
sbml2biopax3.sh (on Linux and MacOS) or sbml2biopax3.bat (on windows) by
passing either a folder that contain SBML files (*.xml) or a single SBML
file and the biopax conversion will be created. If you pass a folder,
every SBML file in it will be converted to biopax.

You can also use a small GUI to run the different converters. The GUI
can be started with the script sbfConverterGUI.sh (on Linux and MacOS)
or sbfConverter.bat (on windows) or by running the "big jar" (java -jar
sbfc-1.3.8-SNAPSHOT-standalone.jar). If using the GUI, the right
converter to select is called "SBML2BioPAX". In the build I have made,
you have three older versions of the converter that I am using to
compare and that will not be present in the future official release, you
can ignore them.



> Secondly, is it possible to include the locus tag information among others into biopax making the RDF model as complete as possible?

Unfortunately, that's not something that the converter is handling at
the moment. First, the information is not really stored in SBML, it
could be stored into some identifiers.org annotations which would be
attached as xref to the BioPAX EntityReference. It could also be present
into custom annotations or notes (like the COBRA specific annotations)
but we don't do anything about them at the moment. Second, looking at
the BioPAX specifications, it is not really clear to me what can be used
to store this information apart from xref.

Thanks,
Nico